A model algorithm based on multilevel flow model (MFM) has been suggested for applying artificial immune network in modeling complex industrial process. Through an analysis of the functions in multilevel flow model, the relative antibody has been designed. Using mass flow and energy flow of multilevel flow model, a stimulation propagation path of artificial immune network is built up. Based on input and output components relations of MFM, a mathematical model of the concentration of antibodies is established, and the failure probability of the node is expressed according to the concentration of antibody. In this paper, a simulation has been conducted and the results show that the algorithm can be used for fault diagnosis of complex industrial process.